Using a custom data frame: ValueError: The truth value of a Series is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all().

[pranavlal]

Hi all,
I have a pandas data frame which I am passing into the backtesting framework.
I have got this frame from a machine learning model. The data frame has a column named "signal". If "signal" is 1, then I buy otherwise I sell.
How do I encode this logic into the next function? I tried the following code and got the below error.
ValueError: The truth value of a Series is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all().

class testModel(Strategy):
def init(self):
super().init()

def next(self):
	if (data.signal==1):
		self.buy(1)
	else:
		self.sell(1)

bt = Backtest(data, testModel, commission=.0002, margin=.05)
stats = bt.run()

The problem is at the line if (data.signal==1):

I have tried enclosing this in parenthesis but that did not help.

I did check if custom data frames can be used and from whatever I have read, the above code should work.
Anyone any ideas on how to fix this?
Pranav

[pranavlal]

Hi all,

I solved the problem. I was referencing a global variable and not the class variable called data hence the problem.

I am now encountering an even more baffeling error message but that is a post for a separate thread.

[gksbelle]

How did u solve the problem?

[pranavlal]

Hi,
I used if self.df.signal[-1]==1:

The self.df is an accessor which gives you access to the data frame.

[kernc]

Using self.data.signal[-1] should run much faster.